1 .\" Copyright (c) 1991, 1993
2 .\" The Regents of the University of California. All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
12 .\" 3. All advertising materials mentioning features or use of this software
13 .\" must display the following acknowledgement:
14 .\" This product includes software developed by the University of
15 .\" California, Berkeley and its contributors.
16 .\" 4. Neither the name of the University nor the names of its contributors
17 .\" may be used to endorse or promote products derived from this software
18 .\" without specific prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 .\" @(#)mmap.2 8.4 (Berkeley) 5/11/95
33 .\" $FreeBSD: src/lib/libc/sys/mmap.2,v 1.22.2.12 2002/02/27 03:40:13 dd Exp $
34 .\" $DragonFly: src/lib/libc/sys/mmap.2,v 1.9 2007/05/17 08:19:00 swildner Exp $
41 .Nd allocate memory, or map files or devices into memory
48 .Fn mmap "void *addr" "size_t len" "int prot" "int flags" "int fd" "off_t offset"
52 function causes the pages starting at
54 and continuing for at most
56 bytes to be mapped from the object described by
58 starting at byte offset
62 is not a multiple of the pagesize, the mapped region may extend past the
64 Any such extension beyond the end of the mapped object will be zero-filled.
68 is non-zero, it is used as a hint to the system.
69 (As a convenience to the system, the actual address of the region may differ
70 from the address supplied.)
73 is zero, an address will be selected by the system.
74 The actual starting address of the region is returned.
77 deletes any previous mapping in the allocated address range.
79 The protections (region accessibility) are specified in the
85 .Bl -tag -width PROT_WRITE -compact
87 Pages may not be accessed.
93 Pages may be executed.
98 parameter specifies the type of the mapped object, mapping options and
99 whether modifications made to the mapped copy of the page are private
100 to the process or are to be shared with other references.
101 Sharing, mapping type and options are specified in the
105 the following values:
106 .Bl -tag -width MAP_HASSEMAPHORE
108 Map anonymous memory not associated with any specific file.
109 The file descriptor used for creating
114 parameter is ignored.
116 .\"Mapped from a regular file or character-special device memory.
118 Do not permit the system to select a different address than the one
120 If the specified address contains other mappings those mappings will
122 If the specified address cannot otherwise be used,
129 must be a multiple of the pagesize.
131 Try to do a fixed mapping but fail if another mapping already exists in
132 the space instead of overwriting the mapping.
134 When used with MAP_STACK this flag allows one MAP_STACK mapping to be
135 made within another (typically the master user stack), as long as
136 no pages have been faulted in the area requested.
137 .It Dv MAP_HASSEMAPHORE
138 Notify the kernel that the region may contain semaphores and that special
139 handling may be necessary.
141 Region is not included in a core file.
143 Causes data dirtied via this VM map to be flushed to physical media
144 only when necessary (usually by the pager) rather than gratuitously.
145 Typically this prevents the update daemons from flushing pages dirtied
146 through such maps and thus allows efficient sharing of memory across
147 unassociated processes using a file-backed shared memory map. Without
148 this option any VM pages you dirty may be flushed to disk every so often
149 (every 30-60 seconds usually) which can create performance problems if you
150 do not need that to occur (such as when you are using shared file-backed
151 mmap regions for IPC purposes). Note that VM/filesystem coherency is
152 maintained whether you use
154 or not. This option is not portable
157 platforms (yet), though some may implement the same behavior
161 Extending a file with
163 thus creating a big hole, and then filling the hole by modifying a shared
165 can lead to severe file fragmentation.
166 In order to avoid such fragmentation you should always pre-allocate the
167 file's backing store by
169 zero's into the newly extended area prior to modifying the area via your
171 The fragmentation problem is especially sensitive to
173 pages, because pages may be flushed to disk in a totally random order.
175 The same applies when using
177 to implement a file-based shared memory store.
178 It is recommended that you create the backing store by
180 zero's to the backing file rather than
183 You can test file fragmentation by observing the KB/t (kilobytes per
184 transfer) results from an
186 while reading a large file sequentially, e.g. using
187 .Dq Li dd if=filename of=/dev/null bs=32k .
191 function will flush all dirty data and metadata associated with a file,
192 including dirty NOSYNC VM data, to physical media. The
196 system call generally do not flush dirty NOSYNC VM data.
199 system call is obsolete since
201 implements a coherent filesystem buffer cache. However, it may be
202 used to associate dirty VM pages with filesystem buffers and thus cause
203 them to be flushed to physical media sooner rather than later.
205 Modifications are private.
207 Modifications are shared.
209 Map the area as a stack.
217 should include at least
222 a memory region that grows to at most
224 bytes in size, starting from the stack top and growing down. The
225 stack top is the starting address returned by the call, plus
228 The bottom of the stack at maximum growth is the starting
229 address returned by the call.
231 The entire area is reserved from the point of view of other
233 calls, even if not faulted in yet.
235 WARNING. We currently allow
237 mappings to provide a hint that points within an existing
239 mapping's space, and this will succeed as long as no page have been
240 faulted in the area specified, but this behavior is no longer supported
241 unless you also specify the
249 is used, you cannot count on the returned address matching the hint
251 .It Dv MAP_VPAGETABLE
252 Memory accessed via this map is not linearly mapped and will be governed
253 by a virtual page table. The base address of the virtual page table may
258 Virtual page tables work with anonymous memory but there
259 is no way to populate the page table so for all intents and purposes
261 can only be used when mapping file descriptors. Since the kernel will
262 update the VPTE_M bit in the virtual page table, the mapping must R+W
263 even though actual access to the memory will be properly governed by
264 the virtual page table.
266 Addressable backing store is limited by the range supported in the virtual
267 page table entries. The kernel may implement a page table abstraction capable
268 of addressing a larger range within the backing store then could otherwise
269 be mapped into memory.
274 function does not unmap pages, see
276 for further information.
278 The current design does not allow a process to specify the location of
280 In the future we may define an additional mapping type,
283 the file descriptor argument specifies a file or device to which swapping
286 Upon successful completion,
288 returns a pointer to the mapped region.
289 Otherwise, a value of
293 is set to indicate the error.
301 was specified as part of the
305 was not open for reading.
310 were specified as part of the
316 was not open for writing.
319 is not a valid open file descriptor.
322 was specified and the
324 parameter was not page aligned, or part of the desired address space
325 resides out of the valid address space for a user process.
331 was specified and the
333 parameter was not -1.
336 has not been specified and
338 did not reference a regular or character special file.
341 was not page-aligned.
347 was specified and the
349 parameter wasn't available.
351 was specified and insufficient memory was available.
352 The system has reached the per-process mmap limit specified in the
367 is limited to 2GB. Mmapping slightly more than 2GB doesn't work, but
368 it is possible to map a window of size (filesize % 2GB) for file sizes
369 of slightly less than 2G, 4GB, 6GB and 8GB.
371 The limit is imposed for a variety of reasons.
372 Most of them have to do
375 not wanting to use 64 bit offsets in the VM system due to
376 the extreme performance penalty.
379 uses 32bit page indexes and
382 a maximum of 8TB filesizes.
383 It's actually bugs in
384 the filesystem code that causes the limit to be further restricted to
385 1TB (loss of precision when doing blockno calculations).
387 Another reason for the 2GB limit is that filesystem metadata can
388 reside at negative offsets.